Think Strategies

import { z } from "zod"; // Export the schema for use in index.js export default z.object({ // Core properties from original schema thought: z.string().describe("Your current thinking step"), nextThoughtNeeded: z.boolean().describe("Whether another thought step is needed"), thoughtNumber: z.number().int().min(1).describe("Current thought number"), totalThoughts: z.number().int().min(1).describe("Estimated total thoughts needed"), isRevision: z.boolean().optional().describe("Whether this revises previous thinking"), revisesThought: z.number().int().min(1).optional().describe("Which thought is being reconsidered"), branchFromThought: z.number().int().min(1).optional().describe("Branching point thought number"), branchId: z.string().optional().describe("Branch identifier"), needsMoreThoughts: z.boolean().optional().describe("If more thoughts are needed"), // Strategy selection and metadata strategy: z.enum([ "linear", "chain_of_thought", "react", "rewoo", "scratchpad", "self_ask", "self_consistency", "step_back", "tree_of_thoughts", "trilemma", "cyclic_reasoning" ]).describe("The thinking strategy being employed"), currentStage: z.string().optional().describe("Current stage in the thinking process flow"), // Strategy-specific properties // ReAct properties action: z.string().optional().describe("Specific action to take for gathering information (ReAct)"), observation: z.string().optional().describe("Result or observation from executing an action (ReAct)"), // ReWOO properties planningPhase: z.boolean().optional().describe("Whether currently in planning phase (ReWOO)"), toolCalls: z.array(z.object({ tool: z.string(), input: z.string() })).optional().describe("Planned tool calls with inputs (ReWOO)"), // Scratchpad properties stateVariables: z.record(z.any()).optional().describe("Current state of variables being tracked (Scratchpad)"), // Self-Ask properties subQuestion: z.string().optional().describe("Current sub-question being addressed (Self-Ask)"), subQuestionAnswer: z.string().optional().describe("Answer to the current sub-question (Self-Ask)"), subQuestionNumber: z.number().int().min(1).optional().describe("Current sub-question number (Self-Ask)"), // Self-Consistency properties reasoningPathId: z.string().optional().describe("Identifier for current reasoning path (Self-Consistency)"), pathAnswers: z.array(z.object({ pathId: z.string(), answer: z.string() })).optional().describe("Answers from different reasoning paths (Self-Consistency)"), // Step-Back properties generalPrinciple: z.string().optional().describe("General principle or approach identified (Step-Back)"), // Tree of Thoughts properties approachId: z.string().optional().describe("Identifier for current approach (ToT)"), approaches: z.array(z.object({ id: z.string(), description: z.string(), promise: z.number().min(0).max(10).optional() })).optional().describe("Different approaches being explored (ToT)"), evaluationScore: z.number().min(0).max(10).optional().describe("Evaluation score for current branch (ToT)"), // Trilemma properties objectives: z.array(z.object({ id: z.string(), name: z.string(), description: z.string(), currentScore: z.number().min(0).max(1), threshold: z.number().min(0).max(1), priority: z.number().min(0).max(1).optional() })).optional().describe("Three competing objectives with scores and thresholds (Trilemma)"), tradeOffMatrix: z.array(z.object({ improving: z.string(), affecting: z.string(), impact: z.number().min(-1).max(1) })).optional().describe("How improving one objective affects others (Trilemma)"), iterationNumber: z.number().int().min(1).optional().describe("Current iteration in the satisficing process (Trilemma)"), equilibriumReached: z.boolean().optional().describe("Whether all objectives meet their thresholds (Trilemma)"), propagatedSolution: z.object({ configuration: z.record(z.number()), overallScore: z.number().min(0).max(1) }).optional().describe("Solution propagated from previous iteration (Trilemma)"), // Cyclic Reasoning properties reasoningApproach: z.enum([ "thought-first", "question-first", "solution-first" ]).optional().describe("The cyclic reasoning approach being used (Cyclic Reasoning)"), currentElement: z.enum([ "thought", "question", "solution" ]).optional().describe("Current element in the reasoning cycle (Cyclic Reasoning)"), cycleNumber: z.number().int().min(1).optional().describe("Current cycle number (Cyclic Reasoning)"), elementOrder: z.array(z.string()).optional().describe("Order of elements for current approach (Cyclic Reasoning)"), domainContext: z.string().optional().describe("Problem domain for approach selection (Cyclic Reasoning)"), approachRationale: z.string().optional().describe("Reason for selecting this approach (Cyclic Reasoning)"), cycleComplete: z.boolean().optional().describe("Whether the current cycle is complete (Cyclic Reasoning)"), needsApproachAdjustment: z.boolean().optional().describe("Whether to switch reasoning approaches (Cyclic Reasoning)"), // Verification and solution properties hypothesis: z.string().optional().describe("Current solution hypothesis"), verificationResult: z.boolean().optional().describe("Whether the hypothesis has been verified"), verificationReasoning: z.string().optional().describe("Reasoning behind verification result"), finalAnswer: z.string().optional().describe("Final verified answer to the problem"), // Semantic routing properties currentState: z.string().optional().describe("Current state in the thinking flow"), stateDescription: z.string().optional().describe("Description of the current state"), sessionToken: z.string().optional().describe("Session identifier token"), availableActions: z.record(z.object({ description: z.string(), requiredInputs: z.array(z.string()).optional(), optionalInputs: z.array(z.string()).optional(), hints: z.record(z.string()).optional(), nextState: z.string().optional(), isGlobal: z.boolean().optional() })).optional().describe("Available actions with semantic hints"), canSwitchStrategy: z.boolean().optional().describe("Whether strategy switching is allowed from current state") });